About

This code has been produced during while writing my Ph.D. (Dr.-Ing.) thesis at the institut of automatic control, RWTH Aachen University. If you find it helpful for your research please cite this:

T. Redick, „Bayesian inference for CAD-based pose estimation on depth images for robotic manipulation“, RWTH Aachen University, 2024. doi: 10.18154/RWTH-2024-04533.

First use

Set up a virtual environment and install BlenderProc via pip:

chmod +x setup_venv.sh
./setup_venv.sh

Now activate the virtual environment and download the required resources:

source venv/bin/activate
blenderproc download cc_textures cc_textures

Generating pose datasets

Run the scripts

./generate_lm.sh
./generate_lmo.sh

Since the steri dataset is not officially part of the BOP datasets, you will have to replace the dataset_params.py of the Blender installation with the provided one. Something like ~/blender/blender-3.5.1-linux-x64/custom-python-packages/lib/python3.10/site-packages/bop_toolkit_lib/dataset_params.py. Reinstall the BOP toolkit:

pip install -e bop_toolkit

Then simply run

./generate_steri.sh